Cocaine addiction remains a serious public health problem. Current treatment methods for cocaine addicts, particularly those seeking treatment in urban settings, are generally inadequate. Thus, there is a significant need to develop and test treatments for cocaine dependence. The current proposal requests continued support for further data analysis of the NIDA Collaborative Cocaine Treatment Study, a large, multi-site psychotherapy clinical trial for patients with DSM-IV cocaine dependence. Five research sites in conjunction with the Coordinating Center, using an identical research protocol, examined the efficiency of two psychotherapies, cognitive-behavioral (CT) and psychodynamic supportive- expressive (SE) psychotherapy, and two drug counseling approaches, individuals, individual drug counseling (ICD) and group drug counseling (GCD). Although the project is fully on schedule, further support is needed to conduct a variety of important additional analyses that are of vital importance to understanding the nature of the efficacy results and pointing the way for future studies. The further analyses involve testing a number of hypotheses about predictors and mediators of treatment response. Three general themes organize the hypotheses for further analysis. These are: (1) Which patients benefit most from psychosocial treatment for cocaine dependence? (2) Do intra-session and extra-session processes influence outcome? And, (3) Do changes in theory-specific constructs mediate change in drug use and other outcomes? In addition to further data analyses, a second general aim of the current proposal is to create a public use dataset that will allow other investigations to pursue questions within our large and rich database.